An End-to-End Ukrainian RAG for Local Deployment. Optimized Hybrid Search and Lightweight Generation

TL;DR

This paper introduces an efficient Ukrainian RAG system with a custom search pipeline and lightweight model, enabling high-quality question answering on limited hardware, achieving 2nd place in a shared task.

Contribution

It presents a novel end-to-end Ukrainian RAG architecture with optimized search and model compression for local deployment.

Findings

Achieved 2nd place in the UNLP 2026 Shared Task.

Demonstrated high-quality QA on resource-constrained hardware.

Developed a specialized Ukrainian language model fine-tuned on synthetic data.

Abstract

This paper presents a highly efficient Retrieval-Augmented Generation (RAG) system built specifically for Ukrainian document question answering, which achieved 2nd place in the UNLP 2026 Shared Task. Our solution features a custom two-stage search pipeline that retrieves relevant document pages, paired with a specialized Ukrainian language model fine-tuned on synthetic data to generate accurate, grounded answers. Finally, we compress the model for lightweight deployment. Evaluated under strict computational limits, our architecture demonstrates that high-quality, verifiable AI question answering can be achieved locally on resource-constrained hardware without sacrificing accuracy.